This document is relevant for: Inf1, Inf2, Trn1, Trn1n

Setup for PyTorch Neuron (Inf1)

• Fresh install
  • Develop on AWS ML accelerator instance
  • Compile on compute instance
  • Deploy on AWS ML accelerator instance
• Update to latest release
  • Develop on AWS ML accelerator instance
  • Compile on compute instance
  • Deploy on AWS ML accelerator instance
• Install previous releases
  • Neuron 2.5.0
    • Develop on AWS ML accelerator instance
    • Compile on compute instance
    • Deploy on AWS ML accelerator instance
  • Neuron 2.4.0
    • Develop on AWS ML accelerator instance
    • Compile on compute instance
    • Deploy on AWS ML accelerator instance
  • Neuron 2.3.0
    • Develop on AWS ML accelerator instance
    • Compile on compute instance
    • Deploy on AWS ML accelerator instance
  • Neuron 1.19.0
    • Develop on AWS ML accelerator instance
    • Compile on compute instance
    • Deploy on AWS ML accelerator instance
  • Neuron 1.18.0
    • Develop on AWS ML accelerator instance
    • Compile on compute instance
    • Deploy on AWS ML accelerator instance
  • Neuron 1.17.2
    • Develop on AWS ML accelerator instance
    • Compile on compute instance
    • Deploy on AWS ML accelerator instance
  • Neuron 1.16.3
    • Develop on AWS ML accelerator instance
    • Compile on compute instance
    • Deploy on AWS ML accelerator instance
  • Neuron 1.16.2
    • Develop on AWS ML accelerator instance
    • Compile on compute instance
    • Deploy on AWS ML accelerator instance
  • Neuron 1.16.1
    • Develop on AWS ML accelerator instance
    • Compile on compute instance
    • Deploy on AWS ML accelerator instance
  • Neuron 1.15.2
    • Develop on AWS ML accelerator instance
    • Compile on compute instance
    • Deploy on AWS ML accelerator instance
  • Neuron 1.15.1
    • Develop on AWS ML accelerator instance
    • Compile on compute instance
    • Deploy on AWS ML accelerator instance
  • Neuron 1.15.0
    • Develop on AWS ML accelerator instance
    • Compile on compute instance
    • Deploy on AWS ML accelerator instance
  • Neuron 1.14.2
    • Develop on AWS ML accelerator instance
    • Compile on compute instance
    • Deploy on AWS ML accelerator instance
• Install with support for cxx11 ABI
  • Installation
  • FAQ
    • When should I use a cxx11 torch-neuron wheel?
    • Can I download a library/header zip file similar to the torch distribution?
    • How can I know which ABI torch-neuron is using?
    • How can I know which ABI torch is using?
  • Troubleshooting
    • What python errors could I see if I mix ABI versions?
    • What compiler/linking errors could I see if I mix ABI versions?

This document is relevant for: Inf1, Inf2, Trn1, Trn1n